U.S. patent application number 15/367866 was filed with the patent office on 2017-06-08 for method for adapting the brightness of an x-ray image, x-ray device, computer program and data medium.
The applicant listed for this patent is Markus Kowarschik, Andreas Maier, Anton Nekovar, Reinhard Stadler. Invention is credited to Markus Kowarschik, Andreas Maier, Anton Nekovar, Reinhard Stadler.
Application Number | 20170161895 15/367866 |
Document ID | / |
Family ID | 58722630 |
Filed Date | 2017-06-08 |
United States Patent
Application |
20170161895 |
Kind Code |
A1 |
Kowarschik; Markus ; et
al. |
June 8, 2017 |
METHOD FOR ADAPTING THE BRIGHTNESS OF AN X-RAY IMAGE, X-RAY DEVICE,
COMPUTER PROGRAM AND DATA MEDIUM
Abstract
A method for adapting the brightness of an X-ray image is
provided. The X-ray image is recorded using a filter attenuating
X-ray radiation used for recording the X-ray image differently in
at least two spatial filter regions. The method includes
determining image regions mapping the filter regions from filter
parameters and recording geometry parameters describing the filter
regions from at least one evaluation line running perpendicular to
a boundary between image regions. The method also includes
determining, for each evaluation line, a correction value
describing a difference in brightness between the image regions
from an image value profile along the evaluation line in an
evaluation area containing the boundary, determining at least one
correction factor from the at least one correction value, and
adapting the brightness between the at least two image regions by
scaling the image values with the correction factor.
Inventors: |
Kowarschik; Markus;
(Nurnberg, DE) ; Maier; Andreas; (Erlangen,
DE) ; Nekovar; Anton; (Neunkirchen, DE) ;
Stadler; Reinhard; (Erlangen, DE) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Kowarschik; Markus
Maier; Andreas
Nekovar; Anton
Stadler; Reinhard |
Nurnberg
Erlangen
Neunkirchen
Erlangen |
|
DE
DE
DE
DE |
|
|
Family ID: |
58722630 |
Appl. No.: |
15/367866 |
Filed: |
December 2, 2016 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06T 2207/10121
20130101; G06T 2207/10116 20130101; G06T 5/008 20130101; G06T
7/0012 20130101 |
International
Class: |
G06T 7/00 20060101
G06T007/00 |
Foreign Application Data
Date |
Code |
Application Number |
Dec 4, 2015 |
DE |
102015224331.1 |
Claims
1. A method for adapting the brightness of an X-ray image recorded
using a filter attenuating, differently in at least two spatial
filter regions, X-ray radiation used for recording the X-ray image,
the method comprising: determining image regions mapping the filter
regions from filter parameters; recording geometry parameters
describing the filter regions from at least one evaluation line;
determining, for each evaluation line, a correction value
describing a difference in brightness between the image regions
from an image value profile along the evaluation line in an
evaluation area containing the boundary; determining at least one
correction factor from the at least one correction value; and
adapting the brightness between the at least two image regions by
scaling the image values with the correction factor.
2. The method of claim 1, wherein the at least one evaluation line
runs along rows, columns, or rows and columns of the X-ray image,
virtual image values along the evaluation line are determined from
the image values of the X-ray image by a coordinate transformation,
a rebinning, or a coordinate transformation and a rebinning, or a
combination thereof.
3. The method of claim 1, further comprising: determining a
plurality of candidate lines along the boundary between two image
regions; and using the plurality of candidate lines, fulfilling a
homogeneity criterion for the image value profile in the evaluation
area for the two image regions, as evaluation lines.
4. The method of claim 3, wherein the homogeneity criterion
describes, at the boundary, a transition from a homogeneous area of
a first image region of the two image regions adjacent to the
boundary to a homogeneous area of a second image region of the two
image regions adjacent to the boundary.
5. The method of claim 3, wherein, within the scope of the
homogeneity criterion, the image value profile in the evaluation
area is fitted to a predefined transition function describing a
transition between homogeneous areas in the two image regions, and
wherein the homogeneity criterion is deemed fulfilled if a quality
value describing the quality of the fit exceeds a limit value.
6. The method of claim 5, wherein a sigmoid function is used as the
transition function, a correlation value is used as the quality
value, or a combination thereof, the transition function, at least
one transition-function parameter describing a concrete form of the
transition function, or a combination thereof is chosen depending
on the X-ray spectrum used for the recording, depending on a
calibration measurement, or depending on a combination thereof, or
a combination thereof.
7. The method of claim 5, wherein the limit value is chosen
depending on the quality values of the candidate lines.
8. The method of claim 1, wherein the adapting comprises adapting
the brightness to a brightness level of a reference image region
assigned to a filter region of minimal attenuation.
9. The method of claim 1, wherein, for each image region to be
adapted, a single correction factor to be applied globally in the
image region is determined.
10. The method of claim 1, wherein, when an image series of X-ray
images is recorded, a correction factor determined for a first
X-ray image with the same setting of the filter is used in a
plurality of X-ray images recorded later, the correction factor
determined for the first X-ray image is taken into consideration in
determining new correction factors for the plurality of X-ray
images, or a combination thereof.
11. The method of claim 5, wherein, to map a blurred brightness
transition between image regions in a transition area along the
boundary between the image regions, a correction factor profile
describing a smooth transition between the correction factors of
the image regions is applied instead of the correction factors of
the image regions.
12. The method of claim 11, wherein, when the image value profile
in the evaluation area is fitted to a predefined transition
function describing a transition between homogeneous areas in the
two image regions, at least one fit information describing at least
one fitted transition function is used for defining the correction
factor profile.
13. The method of claim 1, wherein the at least one evaluation line
runs perpendicular to a boundary between image regions.
14. The method of claim 7, wherein a predefined percentage of the
candidate lines fulfills the homogeneity criterion.
15. The method of claim 9, wherein the single correction factor to
be applied globally in the image region is determined using
statistical observation of the correction values of individual
evaluation lines.
16. An X-ray device comprising: a recording arrangement comprising
an X-ray source and a controller, the X-ray source comprising an
assigned, controllable filter, the controller being configured to:
determine image regions mapping the filter regions from filter
parameters; record geometry parameters describing the filter
regions from at least one evaluation line running perpendicular to
a boundary between image regions; determine, for each evaluation
line, a correction value describing a difference in brightness
between the image regions from an image value profile along the
evaluation line in an evaluation area containing the boundary;
determine at least one correction factor from the at least one
correction value; and adapt the brightness between the at least two
image regions by scaling the image values with the correction
factor.
17. A computer program product comprising a non-transitory
computer-readable storage medium storing a program having
instructions loadable directly into a memory of a programmable
system control unit of a magnetic resonance apparatus, and
executable by the programmable system control unit for the
simultaneous reception of magnetic resonance signals from two or
more slices, the instructions comprising: determining image regions
mapping the filter regions from filter parameters; recording
geometry parameters describing the filter regions from at least one
evaluation line running perpendicular to a boundary between image
regions; determining, for each evaluation line, a correction value
describing a difference in brightness between the image regions
from an image value profile along the evaluation line in an
evaluation area containing the boundary; determining at least one
correction factor from the at least one correction value; and
adapting the brightness between the at least two image regions by
scaling the image values with the correction factor.
18. A non-transitory computer readable medium storing instructions
that, when executed, are operable to: determine image regions
mapping the filter regions from filter parameters; record geometry
parameters describing the filter regions from at least one
evaluation line running perpendicular to a boundary between image
regions; determine, for each evaluation line, a correction value
describing a difference in brightness between the image regions
from an image value profile along the evaluation line in an
evaluation area containing the boundary; determine at least one
correction factor from the at least one correction value; and adapt
the brightness between the at least two image regions by scaling
the image values with the correction factor.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] The present patent document claims the benefit of DE
102015224331.1, filed on Dec. 4, 2015, which is hereby incorporated
by reference in its entirety.
BACKGROUND
[0002] The present embodiments relate to a method for adapting the
brightness of an X-ray image recorded using a filter, with the
filter attenuating, differently in at least two spatial filter
regions, the X-ray radiation used for recording the X-ray image.
The present embodiments also relate to an X-ray device, a computer
program and an electronically readable data medium.
[0003] It is known for medical interventions (e.g., minimally
invasive interventions) to be performed under X-ray monitoring. In
such cases, X-ray images of a recording area, including the
intervention area (e.g., the target area), are recorded with an
X-ray device continuously and/or cyclically during the intervention
and displayed to the at least one person performing the
intervention. X-ray images of this type are frequently referred to
as fluoroscopic images. Instruments, such as a catheter, used
within the scope of the intervention, and/or changes occurring as a
result of the intervention may be observed on such X-ray
images.
[0004] X-ray radiation has an ionizing effect such that a patient
and/or other persons involved in the intervention are exposed,
during the intervention, to an X-ray dose. The X-ray dose should be
kept as low as possible. In order to reduce the X-ray radiation
exposure (e.g., for the patient), it has been proposed that a
filter is connected downstream of the X-ray source of the X-ray
device attenuating the X-ray radiation at least in less relevant
portions of the region recorded (e.g., the field of vision of the
X-ray device). This is based on the idea that there is frequently
only one target region within the region recorded that is relevant
for the observer (e.g., a person performing an intervention). Such
a target region is frequently referred to as a "region of interest"
(ROI), and a corresponding filter may be referred to as an ROI
filter. An exemplary design provides that the filter has a central
(e.g., circular) filter region in which no attenuation of the X-ray
radiation is carried out. This inner filter region is to be
directed at the target region (ROI). The inner filter region is
surrounded by a further filter region that has a fixed attenuation
value such that the filter, in this example, includes a total of
two filter regions in which the X-ray radiation is attenuated
differently (e.g., not at all in the inner filter region and based
on a fixed attenuation value in the outer filter region). Other
forms/designs of such an ROI filter are also conceivable (e.g.,
having other forms of inner filter region and/or a larger number of
different filter regions). A filter of this type may be integrated
in the housing of the X-ray source and has changeable filter
regions (e.g., an inner filter region is movable and/or changeable
in size). This makes it possible to reduce drastically overall the
X-ray dose that the patient and the other persons in the
intervention room are exposed.
[0005] An example of an ROI filter of this type is disclosed in
U.S. Pat. No. 5,278,887. The filter component there, to be used
during fluoroscopy, does not allow for attenuated X-ray radiation
for recording a region of interest (ROI) selected by the physician,
and a high-intensity low-noise X-ray image is therefore produced.
In the regions of the recorded region surrounding the target
region, an attenuated radiation is used that provides a less
intensive, rather noisy image. Here, between the filter regions
(e.g., the inner filter region assigned to the ROI) and the outer
filter region having a fixed attenuation value, a transition region
may be provided in which the thickness of the filter and/or the
attenuation value preferably increase(s) in a linear manner.
[0006] Before fluoroscopic X-ray images are displayed to the at
least one person performing the intervention and/or are otherwise
processed further, the images are usually subjected to image
processing (e.g., that accentuates edges, reduces noise and the
like). The use of the filter gives rise to an X-ray image that has
different brightnesses in the image regions assigned to the
different filter regions, and therefore mapping these filter
regions presents a problem for some image processing algorithms.
Moreover, X-ray images of an even brightness and/or intensity
distribution are easier for medical personnel (e.g., persons
involved in an intervention) to interpret. It is therefore useful
to undertake a brightness adaptation of the X-ray images such that
the X-ray images appear as similar as possible in all image
regions.
[0007] To this end, in U.S. Pat. No. 5,278,887 discusses a
real-time image processing system was proposed that averages values
of the intensities, ultimately averaging the image values, forming,
in the various image regions in order to obtain, based upon these
average values, a brightness adaptation through a linear,
analytically derived transformation. For transition regions, an
approach is proposed that examines the profile within the
transition region in a linearized manner. However, this gives rise
to a plurality of disadvantages. Due to the fact that the average
value is generally formed across all image values in the image
regions, no consideration is given to whether and to what extent an
average identical image value, therefore an average identical
brightness, is expected. Therefore, depending on the anatomy
recorded, brightness differences in the image regions may occur.
This also applies to the transition regions considered separately.
This may lead to X-ray images that are inadequately improved in
terms of their quality.
[0008] Newer X-ray devices, intended to be used within the scope of
medical interventions, often also have functionalities that may
automatically adapt the ROI filter (e.g., with regard to the target
region (ROI)). To this end, it has been proposed that the viewing
direction of a person performing the intervention be determined
with the aid of eye tracking technologies and that the location of
the ROI, and the position the non-attenuating (e.g., the inner,
filter region), be adapted depending on this viewing direction. For
example, a proposal of this type is described in US 2012/0187312.
Frequently changing filter settings, and the different anatomies
located in the focus of attention, further exacerbates the problems
with brightness adaptation as proposed by the prior art, as there
are constantly different situations arising that result in
brightness adaptation of different levels of quality.
SUMMARY AND DESCRIPTION
[0009] The scope of the present invention is defined solely by the
appended claims and is not affected to any degree by the statements
within this summary.
[0010] The present embodiments may obviate one or more of the
drawbacks or limitations in the related art. For example, improved
brightness adaptation in X-ray images recorded with an ROI filter
is provided.
[0011] A method a method for adapting the brightness of an X-ray
image recorded using a filter is provided. The method includes
determining image regions mapping the filter regions from filter
parameters and recording-geometry parameters describing the filter
regions and from at least one evaluation line running (e.g.,
perpendicular to a boundary) between image regions. For each
evaluation line, the method includes determining a correction value
describing the difference in brightness between the image regions
from the image value profile along the evaluation line in an
evaluation area containing the boundary, determining at least one
correction factor from the at least one correction value and
adapting the brightness between the at least two image regions
through scaling of the image values with the correction factor.
[0012] The method may be used in the fluoroscopic monitoring of a
medical intervention (e.g., a minimally invasive intervention), and
a filter with spatially different attenuation (e.g., the ROI filter
described) is therefore used. The filter has different filter
regions where X-ray radiation used for recording the X-ray image is
attenuated differently (e.g., no X-ray radiation in an inner filter
region provided for recording the target region (ROI) and a fixed
attenuation value in at least another filter region surrounding the
inner filter region). The use of the filter (e.g., integrated in
the housing of the X-ray source) has the consequence that the
recorded region (e.g., visible on the X-ray image) of the recorded
object (e.g., the patient) was exposed in at least two subregions.
Each subregion corresponds to a filter region and to different
incident X-ray radiation intensity, leading to different
brightnesses despite equal attenuation by the object in different
image regions of the X-ray image (e.g., the image regions
corresponding to the mappings of the filter regions). Where the
recording geometry is known, the filter properties (e.g., of the
attenuation values) are known and the position/location of the
filter regions is known and the location of the image regions in
the X-ray image may be inferred by making corresponding geometric
observations.
[0013] The filter may be controllable such that the filter regions
may be changed. For example, a non-attenuating filter area may be
directed at the target region (ROI) and may be changeable in
position depending on user inputs and/or tracking of the viewing
direction of a user. Linear motors may be provided as actuators for
this purpose. Because the filter is may be controlled via the
control device of the X-ray device itself, the relevant filter
parameters describing the position of the filter regions are, like
the recording-geometry parameters that describe the recording
geometry, available centrally and may be used to determine the
location of the image regions in the X-ray image corresponding to
the filter regions. Transition regions (e.g., such as have a filter
thickness increasing in a linear manner) may also be filter
regions, but defined adjacent filter regions may be used, each
regions with a fixed attenuation value (possibly even of zero).
[0014] By determining the spatial location of the image regions,
the filter parameters (e.g., describing the current spatial
location of the filter regions) and recording-geometry parameters
may be to assess a hypothesis for the location of the image regions
assigned to the filter regions. The hypothesis is checked and/or
refined based on images. For example, when the filter parameters
and/or the recording-geometry parameters are not available
sufficiently, accurately or reliably, refining the determination of
the position of the image regions based on images may be provided.
For example, after rough positioning in the form of at least one
hypothesis for the filter region, corresponding forms may be laid
over the X-ray image in the position of the hypothesis and a
correlation variable may be determined with the actual image data.
The correlation variable may be used as part of an optimization
method (e.g., moving the form in order to perform fine positioning
or to select from several hypotheses one that has the best
correlation).
[0015] As the basis for the brightness adaptation, the method
includes using evaluation lines (e.g., running essentially
perpendicular to boundaries between image regions). Brightness
profiles, and thus the image data profile, along the evaluation
lines may be evaluated to determine brightness differences in the
image areas. The use of such pinpointed evaluation lines opens up a
plurality of new options for improving the overall image quality
(e.g., assessing transition areas as a suitable determination of
correction values, obtaining additional information about the image
value profile in the transition area between image regions and the
like). More targeted brightness adaptation is provided, based on an
improved database, delivering brightness-adapted X-ray images of
high quality. If additional image-processing algorithms are used
and/or if displayed, the high quality brightness-adapted X-ray
images form a particularly sound basis for further use of the X-ray
images. For example, acceptable, usable X-ray images may be
obtained by persons performing a medical intervention or medical
personnel generally, despite the use of an ROI filter.
[0016] The evaluation area, over which the image data profile is
considered, depending on the application and/or properties of the
X-ray device (e.g., the recording arrangement of the X-ray device),
may be defined as extending on both sides of the boundary for a
predetermined number of pixels, as extending as far as predefined
edge points (e.g., the image edge, image center, and/or region
center) and the like. For example, information about the expected
extent of a transition area between the different brightnesses,
assuming that a homogeneous image area is being considered, may be
factored into the choice of the extent of the evaluation area.
Factors responsible relate to the direction of impingement of the
X-ray radiation on boundaries of filter regions as well as factors
relating to the specific properties of the filter at the
boundaries.
[0017] At least one evaluation line may be chosen as running along
rows and/or columns of the X-ray image, making it easy to identify
the brightness profiles along the evaluation lines. For example,
this type of approach may be used where filter regions are
substantially rectangular in design and their boundaries run in
line with the pixel directions of the X-ray detector used (e.g., a
solid state X-ray detector such as a flat detector having detector
elements distributed matrix-like being used). However, this
approach may also be used for other cases (e.g., the oft-used
circular design of filter regions). Virtual image values along the
evaluation line may be determined from the image values of the
X-ray image by a coordinate transformation and/or a rebinning. The
recorded original image values are resorted and/or rewritten into a
suitable coordinate system such that image value profiles along the
evaluation lines are produced easily. A coordinate system into
which the values are transformed may be deliberately set such that
in a single coordinate transformation suitable virtual image values
is produced for several, or even all, of the evaluation lines to be
used. For example, if a filter with an inner circular filter region
for the target region (ROI) is used, a coordinate system to which
values are transformed may be a polar coordinate system with a
center point corresponding to the center point of the image region
and corresponding to the inner filter region (e.g., as the
evaluation lines run radially out from this center point).
[0018] The implementation of the coordinate transformation and/or
the rebinning is restricted to an application area described by the
extension of the evaluation areas. The coordinate transformation
and/or the rebinning may be limited substantially to the image
areas wherein it is needed, saving calculation time and effort.
[0019] The correction factor may be applied at least in part to the
transformed and/or rebinned image values, whereupon the corrected
image values in the original X-ray image may be determined through
back-transformation. A minimally attenuated image region that is
not to be modified may be loaded with the data of the original
X-ray image (e.g., to refrain from any modifications in an image
region showing the target region (ROI)).
[0020] In order to prevent interpolation artifacts due to
coordinate transformation and/or rebinning, a suitable oversampling
method may be used. For example, oversampling provides a higher
resolution than is ultimately needed for determining the image
value profiles along the evaluation lines. Where a transformation
into polar coordinates is to be made, the angular increment may be
made chosen depending on the maximum radius of the image area to be
transformed (e.g., such that a pixel produced in polar coordinates
substantially matches a pixel in the original coordinate
system).
[0021] A plurality of candidate lines are determined along the
boundary between two image regions, and the candidate lines
fulfilling a homogeneity criterion for the image value profile in
the evaluation area on the part of the two regions used as
evaluation lines. The use of evaluation lines along which
correction values are determined makes it possible to select
evaluation lines from among the available set of evaluation line
such that the selected evaluation lines connect homogenous areas to
one another. The homogeneity criterion describes, at the boundary,
a transition from a homogeneous area adjacent to the boundary of
the one image region to a homogeneous area adjacent to the boundary
of the other image region. In this way, evaluation lines are chosen
that contain the desired information about the brightness
difference factored into or forming the correction value.
[0022] To implement the homogeneity criterion, within the scope of
the homogeneity criterion, the image value profile in the
evaluation area is fitted to a predefined transition function
describing a transition between homogeneous areas in the two image
regions, the homogeneity criterion is fulfilled if a quality value
describing the fit quality exceeds a limit value. Many algorithms
that attempt to adapt a function to a predefined profile already
have inherent values that may be used as quality values (e.g.,
correlation values, regression values and the like). For example, a
sigmoid function that runs in an S-shape may be used as a
transition function. An example of such a sigmoid function is
provided as equation 1:
f(x)=1/(1+e (k*(x+x0))) (Eq. 1)
[0023] In equation 1, the parameter k depends on the X-ray spectrum
used (e.g., and may be derived for example from experiments) and
the parameter x0 is the fit parameter (e.g., and may be determined
by minimizing the least-squares error). The transition function
and/or at least one transition-function parameter describing the
form of the transition function may be chosen depending on the
X-ray spectrum used for the recording and/or a calibration
measurement.
[0024] The remaining deviation of the fitted transition function
from the actual image value profile shows whether the homogeneity
criterion has been fulfilled. For example, a correlation value may
be used as a quality value.
[0025] The limit value may be chosen depending on the quality
values of the candidate lines (e.g., such that a predefined
percentage of the candidate lines fulfills the homogeneity
criterion). For example, a correction value may be determined from
the N % of the image value profiles that have the greatest
correlation coefficient. For example, N may be chosen between 30
and 70% (e.g., 50%). Thus, the underlying data taken as a basis may
be sufficient for determining the correction factor, in that a
predefinable number of evaluation lines is considered as the most
suitable.
[0026] Not all conceivable evaluation lines have to be considered,
and it may be advisable to restrict the number of candidate lines
by predefining an increment along the boundary between image
regions. For candidate lines or evaluation lines emanating from a
center point of a circular image region, a defined angular
increment may be provided, and defining a number of chosen
candidate lines. If no homogeneity criteria are used the evaluation
lines may be selected directly based on an increment along the
boundary.
[0027] The brightness may be adapted to the brightness level of an
excellent reference image region that is assigned to a filter
region of minimal attenuation. If an ROI filter having an inner,
non-attenuating filter region is used (e.g., as a through opening),
the assigned image region may be used as a reference image region
(e.g., as modifications may not be desirable). In the at least one
other image region not corresponding to the reference image region,
the at least one correction factor relating to the brightness or
intensity in the reference image region is applied. For example, if
only two filter regions exist, and therefore also two image
regions, a brightness difference, having to be evened out such that
the correction factor may be determined directly from the
correction values, is ultimately produced relatively directly from
the correction values. For example, if a single evaluation line
describing the transition between homogeneous areas is used, the
ratio and/or the difference of the image values in the respective
homogeneous areas of the image regions may be used as the
correction value. If there are a plurality of evaluation lines,
average values of such ratios may be used.
[0028] For each image region to be adapted, a single correction
factor, to be applied globally in the image region, may be defined
(e.g., through statistical observation of the correction values of
individual evaluation lines). For each image region that is to be
adapted in its brightness (e.g., therefore each image region other
than a reference image region), a global brightness correction
factor is determined and applied to the image values in the
respective image region. However, in a transition area between
image regions, a global brightness correction factor may not be
determined and applied.
[0029] For the individual image regions, a spatially resolved
correction factor may be determined (e.g., a value for the
correction factor may be assigned to each subregion of the image
region). Brightness fluctuations in the image regions, arising due
to the properties of the recording arrangement as a whole or of the
filter, may be managed (e.g., edges are inserted in the image in
adjacent subregions due to the different values of the correction
factor because the different values are undesirable and not
intuitive). In the boundary areas between subregions of different
values of the correction factor, an interpolation is performed to
prevent brightness discontinuities. Suitable interpolation methods
may be used to prevent the introduction of sharp edges between
subregions if possible (e.g., not to introduce new artificial
structures into the image).
[0030] When an image series of X-ray images is recorded, a
correction factor determined for a previous X-ray image with the
same filter setting is used in X-ray images recorded later and/or
is taken into consideration in determining new correction factors
for such X-ray images. The history of correction factors for
adapting brightness may be taken into consideration in order to
improve the brightness adaptation and/or reduce the computational
outlay. For example, if neither the position of the filter regions
of the filter nor the recording geometry between two (e.g.,
fluoroscopic) recordings has changed, the same correction factors
may be used for the corresponding X-ray images. However, newly
acquired information may be used, thus the newly recorded X-ray
image may be used (e.g., in order to improve the determination of
the at least one correction factor statistically by determining
correction values afresh and in the determination of the correction
factors the correction factors of preceding X-ray images recorded
in the same overall configuration). However, findings from
preceding (e.g., fluoroscopic) recording processes may be reused
with regard to changed settings (e.g., a changed position of the
filter regions). For example, a basic expectation in respect of the
brightness difference already exists (e.g., because the attenuation
values of the individual filter regions), irrespective of
positioning of the regions, to remain in many cases the same, even
if filter values are changeable. Through a priori knowledge, a
significant improvement in determining the at least one correction
factor may be achieved in a plurality of cases (e.g., during a
fluoroscopic monitoring process where a plurality of X-ray images
is recorded).
[0031] To map a blurred brightness transition between image regions
in a transition area along the boundary between the image regions,
a correction factor profile describing a smooth transition between
the correction factors of the image regions may be applied instead
of the correction factors of the image regions. This may also be
used when no correction is made in an image region represented by
an appropriate correction factor (e.g., one). Due to the proximity
of the ROI filter to the focal point of the X-ray source, the edge
of a more powerfully attenuating filter region of the filter may
not be mapped as a sharp edge in the X-ray images. Instead, a
profile emerges stemming principally from the geometric properties
of the relevant perspective representation. However, this
implementation recognizes that the filter transition regions
between the filter regions exist having a variable attenuation
value. A wide transition area is produced in the X-ray image.
Because the brightness in the X-ray image no longer changes
abruptly between image regions, image quality may be increased by
creating in the transition area a smooth transition between the
respective correction factors (e.g., one for no multiplicative
correction). Using in a narrow, clearly defined area, a spatially
varying correction factor is used to maximize the image
quality.
[0032] When a fit in a homogeneity criterion is used, at least one
fit information describing at least one fitted transition function
is used for defining the correction factor profile. Thus, the
knowledge already accumulated in the selection of suitable
evaluation lines from candidate lines about how the transition
between the image regions is structured may be used to derive an
appropriate correction factor profile (e.g., providing that in
these areas that the brightness is adapted correctly).
[0033] A discontinuity (e.g., an edge) may also be desired in the
transition area between image regions and may be deliberately
accepted (e.g., where the boundary to an image region,
corresponding to an inner filter region where no attenuation is
taking place, shows the target region of interest (ROI)). Due to
the application of the correction factor without a correction
factor profile, an edge may emerge reproducing the border of the
relevant image region naturally. As such, a selection option may be
provided for the user in an interface (e.g., where the user may
specify whether a smooth transition or an edge-like transition is
desired).
[0034] An application of the present embodiments arises when an ROI
filter is used having an inner, non-attenuating filter region that
is circular and is to be directed toward a target region inside the
recorded region covered by the field of vision. Where a filter
having a circular (e.g., non-attenuating) inner filter region to be
directed toward a target region of the recorded region of the
object to be recorded and an attenuating filter region surrounding
said inner filter region is used for determining the evaluation
lines and the associated image value profiles, a center point of
the image region assigned to the inner filter region is determined
and, to determine the virtual image values for the evaluation lines
extending radially outwardly from this center point, the X-ray
image is transformed into a polar coordinate system. In a polar
coordinate system of this type, the columns correspond to different
angles, and to different candidate lines or evaluation lines, such
that the image value profiles along candidate lines/evaluation
lines may be retrieved and determined easily.
[0035] In addition to the method, the present embodiments include
an X-ray device with a recording arrangement having an X-ray source
with an assigned, controllable filter and a control device
configured for implementing the method discussed above. All
statements in respect of the method discussed above apply to the
X-ray device, such that the advantages of the method may also be
obtained with said X-ray device. For example, filter parameters are
available in the control device describing the current setting of
the filter (e.g., the location of the filter regions and recording
parameters describing the recording geometry). An image region
determining unit of the control device locates image regions
mapping the filter regions. An evaluation line determining unit
selects suitable evaluation lines (e.g., perpendicular to a
boundary to image regions) from candidate lines using a homogeneity
criterion. The evaluation line determining unit may be connected to
a transformation unit that may determine virtual image values along
the evaluation line or candidate line. In a correction value
determining unit of the control device, correction values for the
evaluation lines are determined from the image value profile along
the evaluation line in an evaluation area containing the boundary,
such that a correction unit may determine the at least one
correction factor from the at least one correction value and may
correct the image values correspondingly to adapt the
brightness.
[0036] The present embodiments also provide a computer program
implementing the acts of a method discussed above when the program
is executed on a computing device (e.g., the control device of the
X-ray device). Finally, the present embodiments also provide an
electronically readable data medium on which a computer program as
discussed above is stored, such that when the electronically
readable data medium is read by a computing device, the acts of the
method discussed above are implemented. The data medium may be a
non-transient data medium (e.g., a CD-ROM). The statements made
previously in relation to the method discussed above and in
relation to the X-ray device also continue to apply with regard to
the computer program and the data medium.
BRIEF DESCRIPTION OF THE DRAWINGS
[0037] FIG. 1 shows a schematic diagram of an X-ray device
according to an embodiment.
[0038] FIG. 2 shows an ROI filter used in the X-ray device
according to an embodiment.
[0039] FIG. 3 shows a flow diagram of an exemplary embodiment of a
method.
[0040] FIG. 4 shows image areas in an X-ray image according to an
embodiment.
[0041] FIG. 5 shows an image value profile and a fitted transition
function according to an embodiment.
DETAILED DESCRIPTION
[0042] FIG. 1 shows a schematic diagram of one embodiment of an
X-ray device 1. The X-ray device includes a C-arm 3 mounted on a
stand 2. An X-ray source 4 and an X-ray detector 5 are arranged
opposite each other on the C-arm 3. Using the C-arm 3 and movement
devices of the C-arm 3 (not shown in detail), images of a patient
arranged on a patient bed 6 (e.g., fluoroscopic X-ray images during
a medical intervention) may be recorded from different projection
directions.
[0043] Integrated in the housing of the X-ray source 4, a filter 7
(e.g., an ROI filter 7) allowing different regions in the field of
vision of the X-ray device 1 to be exposed to different radiation
intensities is provided. The field of vision is defined by the
radiation field 8.
[0044] The X-ray detector 5 may be a solid state flat detector
having a plurality of detector elements arranged in matrix-like row
and columns.
[0045] The X-ray device 1 also includes a control device 9
(indicated only schematically) configured for performing a method
according to the one or more of the present embodiments.
[0046] FIG. 2 shows exemplary functioning of the filter 7. A
radiation field 8 emanates from the focal point 10 of the X-ray
source 4 (e.g., the radiation field may cover the entire detector
5). A patient 11 (only indicated), in whom a medical instrument 12
is introduced (e.g., in the form of a catheter), is partially
transilluminated by the X-rays. The radiation field 8 may define
only an entire recording region inside the patient 11 and in high
resolution for the person performing the intervention (e.g., as
this person is interested only in the operating area of the medical
instrument 12, depicted as hatched target region 13). The region
may not necessarily be of interest in its entirety.
[0047] The filter 7 has an attenuating filter element 15. The
filter element 15 is movable via at least one actuator 14 (shown in
FIG. 2 only schematically) controlled by the control device 9 and
having a circular through opening 16 (otherwise having the same
attenuation value throughout).
[0048] In the diagram shown in FIG. 2, the filter element 15 is
positioned by actuator system 14 such that the non-attenuated
X-rays passing through the opening 16 penetrate the target region
13 and are received correspondingly in an image region 17 on the
detector 5.
[0049] The remaining X-ray radiation outside the cone, delimited by
the lines 19, is attenuated by the filter element 15 and strikes
the patient 11 as the object with lower intensity. The X-ray
radiation is received in an outer image area 18 surrounding the
inner round image area 17 by the X-ray detector 5, and a brightness
difference emerges between the image area 17 and the image area 18.
This brightness difference is, prior to image processing of
recorded X-ray images, corrected by the exemplary embodiments
discussed below.
[0050] With regard to FIG. 2, the filter element 15 and the opening
16 may define two filter regions of differing attenuations mapped
onto the corresponding image regions 17, 18.
[0051] The method depicted in FIG. 3 uses the original image data
20 of the recorded X-ray image (e.g., a fluoroscopic image). In act
S1, the filter parameters and recording-geometry parameters present
in the X-ray device (e.g., parameters describing the properties and
the current setting of the filter 7, the recording geometry,
respectively, and other relevant information such as a zoom factor)
are used to locate the image region 17 and to determine the center
point of the image region 17. FIG. 4 shows the circular inner image
region 17, mapping the target region 13, and the image region 18
surrounding the circular inner image region 17 (e.g., appearing
darker due to the attenuation of the X-ray radiation) in the X-ray
image 21. The center point 22 of the image region 17 is shown, as
is the boundary 23 between the image regions 17 and 18.
[0052] If it is not possible to determine the position of the image
area 17 precisely enough from the filter parameter and the
recording-geometry parameters, an image-based refinement may be
made by inserting a circular form at the hypothesized position of
the image region 17 in the X-ray image 21 and checking (e.g., and
if necessary optimizing) the correlation. In the case of a
plurality of hypotheses, the hypothesis with the highest
correlation may be selected. Image processing methods for
determining appropriate correlation variables are already
known.
[0053] In act S2, a coordinate transformation (e.g., effectively a
rebinning of the X-ray image 21 into a polar coordinate system with
an origin in the center point 22 of the image region 17) is
performed. The purpose of the coordinate transformation is that
candidate lines for evaluation lines may be observed (e.g., the
lines standing as perpendicular as possible on the boundary 23).
However, this applies to lines emanating radially from the center
point 22. Two candidate lines 24 that are spaced apart by an
angular increment 25 are indicated in FIG. 4 by way of example. In
the case of image values in the polar coordinate system, the image
value profiles along radial lines 24 correspond exactly to columns
of the correspondingly transformed X-ray image 21, such that image
value profiles may be examined easily.
[0054] To avoid interpolation errors, the angular increment for the
transformation to the polar coordinate system may be chosen
depending on the maximum radius of the transformed image area, and
suitable oversampling may be carried out.
[0055] The whole of the X-ray image 21 may be transformed to the
polar coordinate system, but also may be restricted to an
application area 26, stemming from the fact that the image value
profiles along candidate lines 24 or evaluation lines are analyzed
more precisely only in an evaluation area 27. The evaluation area
27 is broadly defined in the present case, taking into account a
priori knowledge about the imaging behavior of the filter 7 and the
maximum width of transition areas between the image areas 17, 18
(e.g., as a particular number of pixels on either side of the
boundary 23). The application area 26 may be placed such that all
possible evaluation areas 27 are fully contained in the application
area 26.
[0056] In act S3 depicted in FIG. 3, the candidate lines 24 are
determined by selecting candidate lines 24 spaced apart by a
particular angular increment 25. For each of the candidate lines
24, an image value profile in the evaluation area 27 is derived as
a column in the X-ray image 21 transformed into the polar
coordinate system.
[0057] In act S4, evaluation lines are determined from the
candidate lines 24 by checking each of the image value profiles
against a homogeneity criterion. The homogeneity criterion
describes, at the boundary 23, a transition from a homogeneous area
of the one image region 17, 18 adjacent to the boundary 23 to a
homogeneous area of the other image region 18, 17 adjacent to the
boundary 23. To this end (e.g., within the scope of the homogeneity
criterion) the image value profile in the evaluation area 27 is
fitted to a predefined transition function. For example, a sigmoid
function (e.g., equation 1 below) may be used:
f(x)=1/(1+e (k*(x+x0))) (Eq. 1)
[0058] In equation 1, the parameter k is dependent on the X-ray
spectrum used (e.g., measured by appropriate calibration
measurements), and the parameter x0 is the fit parameter (e.g.,
defined such that the least-squares error is minimized).
[0059] An example is shown in FIG. 5. In FIG. 5, the points 28
symbolize the image value profile, and the curve 29 symbolizes the
fitted transition function. The location of the boundary 23 is also
shown. A smooth transition occurs between two substantially
homogeneous areas, such that a high correlation value for the fit
is provided (e.g., indicating that the curve 29 reflects the image
value profile well and that the deviations are small). However, the
corresponding candidate line may not be selected as an evaluation
line because the limit value, which the correlation value is to
exceed as a quality value in the homogeneity criterion, is chosen
dynamically (e.g., such that 50% of the candidate lines 24 are
selected as evaluation lines). This provides that an adequate
number of evaluation lines are available, and that the candidate
lines 24 most likely to be showing a transition between homogeneous
areas candidate lines 24 are selected (e.g., candidate lines 24
with the highest correlation values).
[0060] In act S5, correction values are defined for the evaluation
lines (e.g., as the ratio of the image values in the homogeneous
part of the image region 17 outside the transition area) to the
image values in the homogeneous area of the image region 18.
[0061] The image region 17 (e.g., where non-attenuated radiation is
incident) serves as a reference image region toward which
adaptations are to be made (e.g., the image values are not changed
and correspond to a correction factor of 1. In act S6, a global
correction factor for die image region 18 is determined from the
individual correction values (e.g., constituting local correction
factors, such as through statistical averaging). A weighting may be
applied using the correlation values. In act S7, the image values
in the image region 18 are multiplied with the global correction
factor for the image region 18 to obtain the adaptation.
[0062] To optimize the image quality in the transition area, as
shown in FIG. 5, a smooth progression of the correction factor from
the defined global correction factor to the intended correction
factor of 1 in the image region 17 may be used to achieve the
correction in the transition area. Information on the value profile
in the transition area is already known from the observations in
the homogeneity criterion (e.g., the description by the fitted
transition function, curve 29, being available). This fit
information may be used in order to recreate, in the transition
areas, the smooth transition in accordance with the course of the
curve 29, providing a locally different correction factor to
further improve image quality. This type of local correction in the
transition area is useful, such as in locations where the
transition area extends over a plurality of pixels. The local
correction in the transition area may also be performed on the
transformed virtual image values in the polar coordinate system
(e.g., that are back-transformed so as to produce corrected image
values). The original image data may remain unchanged in the image
area 17 outside the transition area, however, as this data is not
intended to be changed.
[0063] Correlated image data 30 is available. The correlated image
data 30 may be subjected to further image processing and/or may be
displayed to a person performing the intervention.
[0064] Other embodiments in which the image region 18 is subdivided
into subareas (e.g., as arc segments) are provided. Each of the
subareas may be assigned different correction factors, derived from
the local correction values. It is useful to interpolate between
these subregions to provide a smooth transition (e.g., without
sharp edges) based on different correction factors.
[0065] In the case of repeated recording of X-ray images 21 (e.g.,
as part of fluoroscopic monitoring), correction factors of
preceding X-ray images 21 are considered (e.g., used directly) if
the recording geometry and the filter settings have not changed, or
the correction factors of preceding X-ray images are considered at
least in the determination of new correction factors.
[0066] Although the invention has been illustrated and described in
greater detail by the exemplary embodiments, the invention is not
restricted to the disclosed examples. Other variations may be
derived herefrom by one skilled in the art without departing from
the scope of protection of the invention.
[0067] The elements and features recited in the appended claims may
be combined in different ways to produce new claims that likewise
fall within the scope of the present invention. Thus, whereas the
dependent claims appended below depend from only a single
independent or dependent claim, it is to be understood that these
dependent claims may, alternatively, be made to depend in the
alternative from any preceding or following claim, whether
independent or dependent. Such new combinations are to be
understood as forming a part of the present specification.
[0068] While the present invention has been described above by
reference to various embodiments, it should be understood that many
changes and modifications can be made to the described embodiments.
It is therefore intended that the foregoing description be regarded
as illustrative rather than limiting, and that it be understood
that all equivalents and/or combinations of embodiments are
intended to be included in this description.
* * * * *